Hi!

Just FYI: I was adding a MD5PasswordCredential, an interface for it and
a deserializer as well as a serializer to the library this weekend.

How to use it is shown in tests/credentials.txt

I also added an exc.py which holds some exceptions I came up with. Some
need to get more refined to capture all errors accordingly. Some of them
maybe should also go into their respective subpackages, as network etc.

We maybe should also make sure that in cases of errors we always raise
an exception and don't return None or so as an exception is clearer and
can also hold additional data about the exception (at least a reason).

So please look at exc.py and see if there is anything wrong with it,
could be different and so on. If everything is ok I will add calls to
the library (which might also mean changes to interop).

I also started with some refactoring of the message system. It's not
finished and mostly I moved methods around. My goal has been to make it
similar in how the rest of the library works, esp. regarding
serialization. The new syntax now would be:

from message import api
msg = api.new_message('AgentUpdate')
data = api.serialize(msg)

internally it uses the same ISerialization(message) pattern as the rest
of the library does.

I will write more about this tomorrow. If you want to have a look at it
simply go to src/pyogp.lib.base and do

svn switch
https://svn.secondlife.com/svn/linden/projects/2008/pyogp/pyogp.lib.base/branches/
mrtopf-message-refactoring

(all in one line of course).

It's not 100% functional though but I changed
tests/test_template_builder.py in some places to use the new syntax.

We should discuss maybe at tomorrows meeting then. All in all it wasn't
that difficulty as all the methods basically have been there and I just
moved them around a little and adjusted things.
And to say it again: No code has been wasted, it's pretty much still
everything of the original code there.

You can also read about what I did in the commit messages which I tried
to make very verbose.

see http://svn.secondlife.com/trac/linden/timeline

As said, we can discuss this tomorrow and also why I think these changes
are good.

Other places we could work on would be Logging, btw. We should get some
basic logging framework in place so that we don't need to use prints
anymore. These can even stay in place and could be configured by the
application which uses the library.

But more about this another time.

See you tomorrow!

-- Christian






_______________________________________________
Click here to unsubscribe or manage your list subscription:
https://lists.secondlife.com/cgi-bin/mailman/listinfo/pyogp

Reply via email to